To respond to criticisms of the accuracy of its UCR data, the Chicago Police Department assigned four police inspectors to conduct onsite visits and research regarding UCR systems in the New York City, Los Angeles, Philadelphia, Houston, Detroit, and Milwaukee police departments and compare them with the IACP model system. System analysis identified at least 12 decision points where department members could decide not to retain a citizen's report of crime in the system, such as the dispatcher, the police officer, the supervising sergeant, the detective, the State's attorney, and external review. Systems use supervision, bookkeeping, and audit techniques to minimize such reductions. While many commonalities existed among the police departments, attitudes toward the value of UCR statistics varied widely and rules for reporting crimes differed. All departments have difficulties in reconciling UCR definitions with those used in local and State statutes. Some jurisdictions grant great latitude to line personnel in classifying crimes, while others view this task as a final step in the process. The integrity of supervision is presumed in some agencies, but others conduct routine audits. All agencies subscribed to numbering police reports, but only one met the IACP standard that all calls requiring dispatch or assignment of a police officer receive a control number. Chicago's Auditing and Internal Control Division should implement a rigorous auditing program at various stages in the flow of reporting. Chicago has the capability to retrieve basic data on every reported incident and thus does not need to impose a control number for every incident. The Department should publish its guidelines for reporting crime without further delay. Flow charts and individual analyses of the cities' UCR systems are provided.